AI and automation

Introduction and aim

The work package aims to facilitate networking between researchers and clinical staff working with Artificial Intelligence (AI) and automation in all radiotherapy clinics in Denmark, as well as internationally. The work will be linked with the DESIRE data science infrastructure which will also serve as one platform for sharing AI solutions.

As AI is finding its way into all parts of the radiotherapy chain, it is important that we share knowledge and join forces to develop solutions that are robust and can be implemented to serve progress on a national level.

Background

The work package is a continuation of the previous work package “Treatment planning and automation”. One of the primary activities of this WP was to support national workshops for generating consensus guidelines and national data sets for auto-segmentation purposes. ## such workshops were successfully held, and the results are now forming the basis for future work within the current work package.

Automation and use of artificial intelligence is expected to be a cornerstone in technological development of radiotherapy in the coming years. Applications of AI are numerous, with the most mature presently being in image reconstruction and registration, image classification and segmentation, dose optimization and quality assurance. Automating these processes carries a potential to streamline workflows, allow for optimal use of resources, provide consistent and high-quality solution to complex tasks, and pave the way for increased personalization of treatment.

Projects

Current AI/automation research projects of a national/multiinstitutional nature include:

• DAHANCA FORWARD
• HARMONI
• DBCG RT Nation
• DBCG DL Nation

Activities to promote knowledge sharing and collaboration may include among other:

• National workshops on AI in radiotherapy (two have already been held in 2022 and 2023)
• Open source code repositories
• Open networking meetings

WP leaders

Christian Rønn Hansen, Stine Korreman